def test_save_load():
sgnn_fit_data = "oiasdjf opiasd fjopias dfijop asdpoi fjakl;jswe fklzjsnd cvoiuasherf iuhans dfkjuans cvpiouha sdifjn asdkfjn" # TODO: ???.
preproc_sgnn_sklearn_pipeline = get_sgnn_projection_pipeline(
sgnn_training_data=sgnn_fit_data)
sentence_projection_model = make_sentence_model()
test_model_name = MY_MODEL_NAME + ".__unit_test_delete_this_file__"
success = True
try:
# Test save.
save_model(preproc_sgnn_sklearn_pipeline, sentence_projection_model,
test_model_name)
# Test load.
preproc_sgnn_sklearn_pipeline, sentence_projection_model = load_model(
test_model_name)
preproc_sgnn_sklearn_pipeline2, sentence_projection_model2 = load_most_recent_model(
test_model_name, None)
assert isinstance(preproc_sgnn_sklearn_pipeline, BaseEstimator)
assert isinstance(sentence_projection_model, nn.Module)
assert isinstance(preproc_sgnn_sklearn_pipeline2, BaseEstimator)
assert isinstance(sentence_projection_model2, nn.Module)
except:
traceback.print_exc()
success = False
finally:
delete_model(test_model_name)
assert success
def test_model_works():
some_random_d_model = 2 ** 9
five_sentences_of_twenty_words = torch.from_numpy(np.random.random((5, 20, T * d))).float()
five_sentences_of_twenty_words_mask = torch.from_numpy(np.ones((5, 1, 20))).float()
pytorch_model = make_sentence_model(d_model=some_random_d_model, T_sgnn=T, d_sgnn=d)
output_before_match = pytorch_model(five_sentences_of_twenty_words, five_sentences_of_twenty_words_mask)
print(output_before_match.shape)
assert output_before_match.shape == (5, some_random_d_model)
def test_trainer_model_works():
raw_data = "This is a sentence. And this is another one. This test needs sentences as such"
category_per_sentence = [0, 1, 3]
preproc_sgnn_sklearn_pipeline = get_sgnn_projection_pipeline(sgnn_training_data=raw_data)
sentence_projection_model = make_sentence_model(d_ff=1024)
list_of_sentences = raw_data.split(". ")
projected_words = preproc_sgnn_sklearn_pipeline.transform(list_of_sentences)
projected_words, mask = pad_right(projected_words)
target_diagonal_block_matrix = categories_to_block_matrix(category_per_sentence)
model = TrainerModel(sentence_projection_model)
loss = model(projected_words, mask, target_diagonal_block_matrix)
assert type(loss.item()) == float
def test_trainer_model_works_on_gpu():
device_id = 0
torch.cuda.set_device(device_id)
with torch.cuda.device(device_id) as cuda:
raw_data = "This is a sentence. And this is another one. This test needs sentences as such"
category_per_sentence = [0, 1, 3]
preproc_sgnn_sklearn_pipeline = get_sgnn_projection_pipeline(
sgnn_training_data=raw_data)
sentence_projection_model = make_sentence_model(d_ff=1024).cuda(device_id)
list_of_sentences = raw_data.split(". ")
projected_words = preproc_sgnn_sklearn_pipeline.transform(list_of_sentences)
projected_words, mask = pad_right(projected_words)
projected_words, mask = projected_words.cuda(device_id), mask.cuda(device_id)
target_diagonal_block_matrix = categories_to_block_matrix(category_per_sentence).cuda(device_id)
model = TrainerModel(sentence_projection_model).cuda(device_id)
loss = model(projected_words, mask, target_diagonal_block_matrix)
assert loss.is_cuda
def test_model_works_on_gpu():
device_id = 0
with torch.cuda.device(device_id) as cuda:
some_random_d_model = 2 ** 9
five_sentences_of_twenty_words = torch.from_numpy(np.random.random((5, 20, T * d))).float()
five_sentences_of_twenty_words_mask = torch.from_numpy(np.ones((5, 1, 20))).float()
pytorch_model = make_sentence_model(d_model=some_random_d_model, T_sgnn=T, d_sgnn=d)
five_sentences_of_twenty_words = five_sentences_of_twenty_words.cuda(device_id)
five_sentences_of_twenty_words_mask = five_sentences_of_twenty_words_mask.cuda(device_id)
print(type(five_sentences_of_twenty_words), type(five_sentences_of_twenty_words_mask))
print(five_sentences_of_twenty_words.is_cuda, five_sentences_of_twenty_words_mask.is_cuda)
pytorch_model = pytorch_model.cuda(device_id)
output_before_match = pytorch_model(five_sentences_of_twenty_words, five_sentences_of_twenty_words_mask)
assert output_before_match.shape == (5, some_random_d_model)
print(type(output_before_match))
print(output_before_match.is_cuda, output_before_match.get_device())
assert output_before_match.is_cuda
assert five_sentences_of_twenty_words.is_cuda
assert five_sentences_of_twenty_words_mask.is_cuda
def test_can_create_model():
pytorch_model = make_sentence_model()
assert isinstance(pytorch_model, torch.nn.Module)